comp.lang.java.programmer Google Group
Programming in the Java language.

• Re: A question related to type casting
  Whatever type T is.
  using System;
  using System.Collections.Generic;
  namespace E
  {
  public class BigG<T> where T : new()
  {
  public static T GiveMe()
  {
  return new T();
  }
  }
  public class Program
  {
  private static void Analyze(string prefix, Type t)
• Re: Standalone gnu-diff like tool in Java
  It only supports some Windows versions (2000, XP and 2003 - not 9x and
  Vista - and most likely not 2008 when it does not support Vista).
  Infrequent updates (latest version is form 2006).
  Most likely less utilities (my cygwin install has 1805 exe's
  in bin dir).
  Arne
• Re: A question related to type casting
  If developers ignore compiler warnings, then I don't think they
  should complain over runtime error.
  I do.
  Well - a nice yellow warning sign with a warning text similar
  to what -Xlint gives.
  It obviously does.
  But in most cases developers would agree about what is weird.
  It looks very weird to me. The code say "give me something type
• Re: A question related to type casting
  Well - It is a fact that the feature is used in C#.
  I find it difficult to believe that Java programmers would
  not use it if it was available in Java.
  Constructors can be different, but a lot of classes have noarg
  constructors.
  Arne
• Re: A question related to type casting
  I don't agree. Creating a new instance is not something it really
  makes sense to be able to do "generically". Different classes have
  different constructors, with various arguments and different
  requirements. And there's no way to specify a constructor for all
  subclasses with inheritance. Nor would you really want to. A subclass
• Re: A question related to type casting
  Lew wrote:
  WTF?
  I *did* post the warnings that javac gives. Those *are* the warnings
  about the use of unparameterized types.
  Those would be the warnings about the use of unparameterized types.
  Come on, man.
  I wonder why Eclipse chose to go with different warnings from what
  other Java compilers issue. I don't know what Eclipse says; I
• Re: Java6: When are BROWSE and MAIL actions supported on Linux?
  Bernd Eggink schrieb:
  OK, I found it, by trial and error. The 'avahi' library from gnome was
  missing.
  I also found this "technical article":
  [link]
  which says: "On the Solaris Operating System and the Linux platform,
  this API is dependent on Gnome libraries". That's it. Not an overly
• Re: visibility on array elements
  Thanks a lot to you and Arne (and the others that will jump in)
  for answering my question,
  Driss
• Re: A question related to type casting
  I guess that depends on your definition of "just fine". The fact is, the
  compiler passes the code, and the run-time executes it. I prefer no
  warnings when I compile, but they aren't a restriction on the code itself,
  and I'm not seeing the warning you describe anyway.
  I don't get that warning from Eclipse.
• Re: A question related to type casting
  Precisely!
  You're quite right. I hadn't thought this through properly - a F<? extends
  T> is *actually* a F<U extends T>, for some U. You don't know it at
  compile-time, but it's there, and in a reified system, you'd know it at
  runtime.
  tom
• Re: A question related to type casting
  Lew wrote:
  This absolutely does *not* compile "just fine". You get an "unchecked
  cast" warning for the cast to a raw List.
  You're really not supposed to mix raw and generic types in Java. This
  is what Arne was referring to when he spoke of "weird looking code"
  that lets you "shoot yourself in the foot".
• Re: (macosX,netbeans) jndi : can't find serialContextProvider
  there are news: I created a new domain via the console, and I started
  the domain.
  now I have the message : moi.un.plus_2bBean not found.
  NB: the lookup method contains the argument "moi.un.plus_2bBean" which
  is the argument ,in the bean code, of the annotation
  "mappedName("moi.un.plus_2bBea n")".
  olivier.
• Re: A question related to type casting
  On Wed, 31 Dec 2008 05:42:57 -0700, Tom Anderson <t...@urchin.earth.li>
  wrote:
  Impossible to say until you tell me the type parameter. But, for example:
  Cache<String> cache = new Cache<String>();
  Object obj = cache.makeNewElement();
  The type of the instance referred to by "obj" would be "String".
• Re: visibility on array elements
  Yes, you'll get 3L (assuming nothing else has disturbed
  ref[0] in the meantime).
  You can synchronize all accesses on the array itself, as
  above, in which case no accesses will overlap and modifications
  made by any thread will be visible to all others. Note that
  this may create more contention than if you could somehow
• Re: HTML Correctness and Validators
  True.
  For CSS to be a good solution then the HTML need to be designed
  for usage with CSS.
  Arne